Introduction
The main aim is predicting breast cancer patients chance of survival.
- Clean the data
- Augment the data
- Create some plots
- Statistical analysis
- Create the prediction tool
Materials and methods
Materials
We are working with a dataset about Breast Cancer that we have obtained from kaggle website
This is the dataset we are working with:
| patient_id | gender | education | treatment_data | id_healthcenter | id_treatment_region | hereditary_history | birth_date | age | weight |
|---|---|---|---|---|---|---|---|---|---|
| 111036008041 | 0 | 4 | 2019 | 1.11e+09 | 1.11e+09 | 1 | 1989 | 30 | 69 |
| 111035996130 | 0 | 6 | 2019 | 1.11e+09 | 1.11e+09 | 0 | 1989 | 30 | 71 |
| 111035971333 | 0 | 5 | 2019 | 1.11e+09 | 1.11e+09 | 0 | 1989 | 30 | 74 |
| 111036018485 | 0 | 5 | 2019 | 1.11e+09 | 1.11e+09 | 1 | 1989 | 30 | 75 |
| 111035985474 | 0 | 1 | 2019 | 1.11e+09 | 1.11e+09 | 0 | 2009 | 10 | 70 |
| 111035903616 | 0 | 3 | 2019 | 1.11e+09 | 1.11e+09 | 1 | 1989 | 30 | 79 |
Cleaning the data
Cleaning the data
| BEFORE | AFTER |
|---|---|
| THE COLUMNS ARE DIFFERENT TYPES | EACH COLUMN HAS A CORRECT TYPE |
| 0, 1, 2 VALUES | BOLEAN VARIABLES |
| NAMES WITH /R/N | CLEAN NAMES |
| BIRTH DATE WITH 3 CHARACTERS | BIRTH DATE WITH 4 CHARACTERS |
| BLOOD TYPE 44 | CORRECT BLOOD TYPES ONLY |
| WEIRD WEIGHT/AGE CORRELATIONS | ELIMINATING PEOPLE UNDER 20 YEARS OLD AND 35 KG |
| WOMEN AND MEN | ONLY WOMEN |
Augmenting the data
- We have added more informative columns
- We have changed the type of the columns
Statistical analysis
We have created some plots in order to fully understand the data and we have done some statistical analysis like MCA analysis. The plots are shown in the following point: “Results”
Results
Barchar of categorical variables
//: # Variables that affect health (medicines, vicious habits) have a great incidence in breast cancer patients //: # Early menstrual periods before age 12 and starting menopause after age 55 expose women to hormones longer, raising their risk of getting breast cancer
Boxplots stratified on conditions
Boxplots of numerical variables
Histograms stratified on condition
Histograms stratified on condition
//: # In most cases, when having taking medicine the death is higher (no sense). //: # Not drinking alcohol or smoking improves recovery. //: # When taking alcohol and smoking the death is lower (it doesn’t make any sense) //: # These are absolute values, maybe we should calculate some relative values
Densitograms stratified on condition
Histograms stratified on condition
Heatmap of numerical variables
Dynamic representation of variables
MCA analysis
Discussion
Discussion
Conclusion
Conclusion
We have reached the following conclusions
Bibliography
Bibliography
- Breast cancer incidence (invasive) statistics (10 March 2020). Cancer Research UK
- Height Percentile Calculator for Men and Women in the United States. DQYDJ
- Weight Gain After Breast Cancer Diagnosis May Be a Bigger Issue Than Thought in Australia (13 March 2020) BreastCancer.org.
THANKS
FOR YOUR ATTENTION